Various automated methods are known for selecting products based on customer preferences. In one approach, a customer is provided with a desired product by receiving the customer""s preferences and comparing them with the characteristics of discrete products. For example, U.S. Pat. No. 5,056,029 to Cannon describes a method and apparatus for vending social expression cards by querying the customer to input a series of special occasion parameters, searching a database for card designs having special occasion parameters matching those input by the customer, and displaying the matching card designs.
In another approach, customer preferences and discrete product attributes are quantitatively correlated, leading to the suggestion of products having attributes that do not necessarily perfectly match the customer preferences. For example, U.S. Pat. No. 5,550,746 describes a method and apparatus for selecting social expression cards by quantitatively correlating product design characteristics with customer-entered selection criteria.
Both of the above approaches are useful for selecting from among combinations of discrete product attributes. However, some products cannot be characterized as combinations of discrete product attributes. For example, it would be desirable for a vendor to offer a multi-component composition in which amounts of components may be varied continuously, and in which the properties of the composition vary continuously with component amount variations. Such an approach would allow a customer to select a composition optimized for the customer""s particular needs, rather than being restricted to selecting from a few discrete compositions.
There is therefore a need for an automated method and apparatus for specifying a multi-component composition based on customer-specified limitations on its properties.
The above-described and other drawbacks and disadvantages of the prior art are alleviated by a method for specifying a multi-component composition, comprising: receiving, from a user system, a first limitation on a first physical property and a second limitation on a second physical property; determining a recommended composition by searching a first continuous surface relating the first physical property to at least two compositional variables to determine first compositions that satisfy the first limitation, searching a second continuous surface relating the second physical property to at least two compositional variables to determine second compositions that satisfy the second limitation, determining an intersection of the first compositions and the second compositions, and selecting the recommended composition from the intersection of the first compositions and the second compositions; predicting a value of the first physical property and a value of the second physical property for the recommended composition; and presenting, to the user system, the value of the first physical property and the value of the second physical property for the recommended composition.
Other embodiments, including a system for specifying a multi-component composition, a storage medium encoded with machine-readable computer program code for determining a multi-component composition, and a computer data signal for determining a multi-component composition, are described below.